Nudge Security now discovers shadow AI agents through browser activity in addition to direct API integrations with agentic platform providers.
Every AI agent security tool on the market shares the same gap right now. It's not a roadmap problem. It's an architecture problem. And today, we just worked around it.
Nudge Security is now the first and only AI security solution of its kind to discover shadow AI agents without relying on endpoint agents or API integrations alone. With our new browser-based AI agent discovery—live now as part of our AI agent discovery research preview—Nudge Security can see beyond API limitations to discover shadow agents across more agentic platforms, including Cursor automations, Retool, Zapier Agents, Atlassian Rovo, ChatGPT workspace agents, and more.
AI agents are becoming a new layer of business logic. They hold persistent permissions. They connect to corporate apps and data. They execute autonomously. And your employees are building them everywhere—fast.
Here's the catch: the platforms where they're building agents aren't just the tools with enterprise security controls and mature public APIs. They're also the fast, low-friction tools your engineers, ops teams, and PMs already love—and there’s no guarantee that security teams can audit, derisk and govern all of the agentic activities going on.
Reliance on a single discovery and governance method—API—means every other AI agent security tool is structurally bounded by what agentic platform vendors choose to expose. The agents your team is actually building may stay invisible. Until today.
The Nudge Security browser extension—the same one many of our customers already deploy for AI and SaaS discovery and governance—now passively observes when an employee views, lists, or creates an AI agent in a supported platform. The agent flows automatically into your AI agents inventory, mapped to its human creator, with permissions, components, and risk insights attached.
And because the extension is already deployed across most of our customer base, there's nothing new for customers to roll out. If you’re already enrolled the the research preview, just toggle on AI agent discovery in your browser extension settings, and discovery will start automatically.
Browser-based discovery is live now for:
Combine that with our existing API-based discovery—Salesforce Agentforce, Microsoft Copilot, ServiceNow, Gemini, n8n, Tines, Abacus, Workato, and ChatGPT—and you get 17+ platforms across two complementary discovery channels.
Every discovered agent shows up in your AI agents inventory under Identities. You see who built it, what it's connected to, what permissions it holds, and the risks it carries: publicly accessible agents, hardcoded credentials, unauthenticated MCP connections, agents tied to high-risk apps, agents whose creators have left the organization.
From there, you set approval status, assign a technical contact, and nudge the creator directly to confirm intent and remediate when something looks off.
It's the discovery, accountability, and governance loop your AI agent program needs urgently—without waiting on agentic platform providers.
If you're already in our AI agent discovery research preview and the Nudge Security browser extension is deployed, head to your browser extension settings and toggle browser-based agent discovery on. Agents will start populating in inventory as soon as your team views or creates them.
New to Nudge Security? Start a 14-day free trial to get started with AI security and governance.
